Pricing

Self-host Cycles for free. Pay for production assurance when runtime authority becomes part of your agent infrastructure. No license key, no seat fees, and no percentage of agent spend.

Today, paid plans support self-hosted Cycles deployments. A managed cloud (runcycles.ai) is planned but not yet available.

How pricing works

Apache 2.0 open-source software. Cycles is Apache 2.0 open-source software. Self-host it for free. Paid plans do not unlock hidden features; they add support, implementation help, and assurance. No license key, no feature gating, no seat fees.
No agent-spend tax. We never take a percentage of your model, tool, or agent spend. Support pricing is flat and independent of tokens, requests, agents, tenants, or downstream provider spend — the opposite of gateways that bill a cut of every call.
You pay for the production layer, not the software. The software stays open. Paid plans add response targets, upgrade guidance, incident support, architecture review, and audit-ready evidence configuration for teams running Cycles as a control point in their agent infrastructure.

For regulated or enterprise deployments, paid support can include evidence retention, signed CyclesEvidence exports, audit-readiness review, and control-narrative support — see the Compliance Evidence Package below.

What support includes

Cycles runtime / admin / events deployment guidance
Protocol behavior and configuration support
Upgrade and version-pinning guidance
Scope, budget, policy, and evidence-model review
Integration guidance for supported frameworks

What it does not include

Legal advice
Building your agent product
Owning your production operations
Custom feature development unless separately scoped

A workload means one agreed production path where Cycles gates model calls, tool calls, agent runs, or another consequential action flow.

Community

FreeSelf-hosted, Apache 2.0

Evaluating or running non-critical workloads

GitHub issues, docs, and community
Best-effort response
Public security advisories
Public roadmap

Start self-hosting →

Recommended

Production Support

Introductory pricing

$1,500/mo, billed annually$18,000/yr

One production deployment or workload path

One private Slack channel + email
Next-business-day response; SEV-1 target 4 business hours
Early security advisories + upgrade guidance
One annual scope/deployment review
Roadmap input

Talk to the team →

Enterprise

From $6,000/mo, billed annuallyfrom $72,000/yr

Mission-critical, regulated, or multiple workloads

Named engineer + quarterly architecture reviews
Custom incident-response terms, including optional 24×7 SEV-1 coverage
Evidence/retention configuration + auditor Q&A support
Version-pinning + upgrade guidance
Pre-release access for protocol changes
Private Slack, email, and scheduled calls
Prioritized protocol requests

Contact sales →

Enterprise pricing scales with workloads, environments, response coverage, evidence requirements, and 24×7 terms.

Fixed-scope services

One-time engagements with a fixed price and a concrete deliverable. Subscriptions are billed annually; services are 50% on start / 50% on delivery, or net-30.

Production Readiness Review

$7,500~1 week

Architecture, security, and scope-model review plus a Redis HA/capacity plan. Deliverable: written findings, deployment risk register, runtime/admin hardening checklist, recommended policy/scope model, and a 30-day implementation plan.

Integration Sprint

$15,0002 weeks

Wire Cycles into one agreed workload path with reserve / commit / release, shadow mode where applicable, and a working PR or patch set (Python, TypeScript, Spring, or MCP). You provide access to the target codebase, the deployment environment, and a technical owner.

Compliance Evidence Package

$12,000fixed scope

Map Cycles-generated evidence and runtime controls to selected control narratives for EU AI Act readiness, NIST AI RMF, and ISO/IEC 42001. Configure CyclesEvidence signing and retention/cold export, and deliver an auditor-ready evidence pack. This is not legal advice.

Team Enablement Workshop

$3,500half-day · $6,000 full-day

Remote workshop for up to 12 engineers on runtime authority, budget and scope design, and the incident patterns Cycles is built to prevent.

Design partners are a separate track

The Design Partner program is not a pricing tier. It's a small, selective, free cohort: founder access and priority support during a 60-day integration window, in exchange for running a real workload, pushing us hard on what's missing, and an opt-in public artifact. Paid support is for teams that want guaranteed response and expert help without the design-partner commitments — or who are outside or past the cohort.

Running a real multi-tenant workload and willing to shape the next two protocol releases? Apply to become a design partner →

Managed cloud — coming

A fully managed Cycles (runcycles.ai) is on the roadmap: we run Redis, the runtime, the admin server, and the events service, on a SOC 2 track, so you don't have to. It is planned but not yet available — until then, every paid plan above supports your self-hosted deployment.

Join the managed-cloud waitlist →

Frequently asked

Is the software really free? Yes — Apache 2.0, no feature gating, no license key. Paid plans add support and assurance, not capabilities.

Do you take a cut of our agent spend? No. We never take a percentage of your model, tool, or agent spend. Support pricing is flat and independent of usage.

Do I need a support plan to run in production? No. Many teams run self-hosted on community support. Paid plans are for guaranteed response and expert help. Note that Production support covers the Cycles deployment — not building or operating your agent application.

How is a support plan different from the design partner program? The design partner program is free, selective, and feedback-for-access: you run a real workload and shape the roadmap. A support plan is open to anyone, gives you a defined response target, and carries no feedback or co-marketing obligation.

Can a services engagement turn into a subscription? Yes. A Production Readiness Review or an Integration Sprint commonly precedes a Production support plan.

What about managed cloud? It's on the roadmap. Join the waitlist.

How do you bill? Subscriptions are billed annually. Services are 50% on start and 50% on delivery, or net-30.

Talk to the team building Cycles

Most teams reach out for one of four reasons — they all land in the same inbox, and a founder reads every one.

Piloting Cycles on a real stackYou're putting Cycles in front of a real workload — Python, TypeScript, Spring Boot, Rust, MCP — and want implementation help or a second pair of eyes on the integration.
Hit a failure mode we haven't documentedRunaway agent, retry storm, unit mismatch, scope-derivation edge case. Tell us what broke; we'll add it to the runbook.
Feedback on the protocol or roadmapYou've read the spec and have opinions on the error contract, permission model, or something you need in v0.1.26.
Evaluating Cycles for productionYou're evaluating Cycles for production — comparing it to rate limiters, LLM gateways, or observability tools. We'll walk through the spec, demo the runtime, and answer integration questions.Start with the evaluation guide →